This project aims to develop an AI-powered vehicle visual defect detection system leveraging advanced computer vision techniques to automate and enhance traditional manual inspections on manufacturing lines. Utilizing the comprehensive CarDD dataset, the solution employs YOLOv11 with Oriented Bounding Box (OBB) technology to detect and classify various visual defects such as scratches, dents, and paint anomalies. Data augmentation strategies are used to address class imbalances and enhance model robustness. The primary evaluation metric is mean Average Precision (mAP@0.5:0.95 IoU threshold), targeting a minimum accuracy of 80%. Expected outcomes include improved detection accuracy, increased operational efficiency, and reduced inspection costs, offering significant commercial value in automotive manufacturing and quality assurance processes.
Keywords: computer vision, YOLOv11, vehicle defect detection, Oriented Bounding Box (OBB), manufacturing quality control, CarDD dataset
Watch the team present this project at 2:35 in the session recording here.
Faculty Advisor
PhD in Theoretical Physics, MS in Computer Science. Currently – Applied Scientist at Amazon. Previous jobs: Computational Scientist at the Argonne National Laboratory, Scientist at LIGO project of California Institute of Technology. I specialize in scientific computing, HPC, machine learning, data analysis.